Introduction

The onset of the Industrial Revolution has drastically transformed cities worldwide, driven by economic development and globalization. Urbanization, a global trend, is anticipated to surge, constituting 70% of the world’s population by 2050 (Sodiq et al. 2019). China, as a leading developing country, has been actively promoting urbanization through its policies and economic growth. Between 1978 and 2019, the urbanization rate in China experienced substantial growth, rising from 17.9% to 60.6% (NBS 2020). Throughout this period, the Chinese government emerged as a pivotal force, actively steering the trajectory of government-led urbanization. However, negative impacts from urban development, such as environmental pollution, social injustice, and wasteful resource utilization, are evident from the experiences of other developed nations (Wang et al. 2014). This has led urban planners to focus on sustainable cities, which take into consideration sustainable goals such as energy conservation, energy efficiency, human capital development, building standards, and food waste management (Sodiq et al. 2019). Despite the Chinese government’s efforts to promote urbanization, it has also resulted in numerous problems, including a crowded urban populace, employment difficulties, and wealth disparity, leading to a development crisis (Gries and Grundmann 2018; Zheng et al. 2014). To ensure the sustainability of each city, the Chinese government has formulated sustainable development plans, which are essential in promoting sustainable urbanization. The “National New-type Urbanization Plan (2014–2020),” introduced in March 2014, is an example of China’s commitment to sustainable urbanization. This transformative plan signifies a shift from land-centered to people-centered urbanization, emphasizing a “human-centered and environmentally friendly path” for the upcoming decade (Chen et al. 2018; Taylor 2015). In addition to general and short-term policies, the National New-type Urbanization Plan is a nationwide, long-term, and far-reaching policy that warrants special attention for in-depth study. It establishes new developmental goals for China’s robust urbanization process, outlining four primary objectives, including the enhancement of city sustainability (Chu 2020). Government policies and regulations exert a substantial impact on efficiency, exemplified by the influence of environmental regulatory policies on energy consumption efficiency (Chen and Gong 2017). Additionally, initiatives such as the Low Carbon City Pilot Project contribute to the overall improvement of carbon emission efficiency in pilot cities (Fu et al. 2021), while support policies for resource-depleted cities have significantly enhanced energy efficiency (Yu et al. 2022). By examining relative changes in cities within the context of the National New-Type Urbanization Plan, the external situation created by the policy becomes apparent. However, with the advancement of China’s industrialization and urbanization, the focus of urban sustainability must extend beyond environmental assessment.

Effective governance is pivotal in steering urban development towards sustainability, as policies encompassing economic development, social welfare, and environmental protection profoundly influence urban sustainability. However, governance faces impediments such as centralization of power, political risks, and corruption, which can hinder the implementation of sustainable development policies (Neirotti et al. 2014). In tandem with governance, the integration of Information and Communication Technology (ICT) emerges as a transformative force in modern urban development (Odendaal 2003). As such, ICT presents an opportunity to address the problems of urbanization and is an effective means of realizing sustainable urban development and building eco-efficient smart cities. As the second-largest global economy, China has actively pursued the development of smart cities through substantial investments in ICT infrastructure. The information and communication technology (ICT) revolution has made cities more digitized and information-driven, with AI, digital currency and other services becoming a part of daily life (Yao et al. 2020; Yin et al. 2015). The government’s focus on ICT aligns with global trends in smart city development, which leverage technology to address urban challenges and enhance overall urban well-being (Chourabi et al. 2012). Sustainable development, encompassing economic growth, environmental sustainability, and social inclusion, stands as the overarching goal in contemporary urban planning (Roseland 2000). The choice of governance, ICT, and sustainable development as the evaluation framework is rooted in their synergistic impact on urban sustainability. Effective governance is essential for creating and implementing policies that foster sustainability. ICT acts as an enabler, facilitating efficient resource management, data-driven decision-making, and smart solutions for urban challenges. Sustainable development, with its triple bottom line approach, provides a comprehensive lens to evaluate the economic, social, and environmental facets of urban systems.

Data Envelopment Analysis (DEA) is a non-parametric method assessing the efficiency of decision-making units (DMUs) through multiple inputs and outputs, enabling a quantifiable evaluation of a city’s sustainability and efficiency. The literature suggests diverse indicator systems for sustainable cities, constructed with various DEA models. While common indicators like air quality and economic development exist, specific ones such as carbon emission reduction and energy efficiency have gained attention (Sheng et al. 2021; Hu and Wang 2006). Traditionally, traditional DEA models treated DMUs as “black boxes,” focusing on slack in input or output spaces (Suzuki and Nijkamp 2017). However, network DEA breaks down internal organizational efficiency, evaluating both system and process efficiency (Kao 2009; Yang 2016). Despite the widespread use of traditional DEA models, the advanced network DEA approach is underutilized. For instance, Suzuki and Nijkamp (2017) employed advanced CCR and context-dependent (CD) models to rank the efficiency of 35 major global cities. Similarly, Yan et al. (2018) used CCR and BCC models to assess the sustainable development performance of 287 cities in China. The prevailing urban operational efficiency indicator system, utilizing the DEA method, especially within the Chinese context, exposes significant deficiencies. This study proactively closes this gap by introducing a comprehensive evaluation framework that systematically incorporates urban governance, ICT, and sustainability efficiency. Surpassing the constraints of prior models that predominantly focused on individual dimensions, the research propounds a more holistic perspective on urban operations, thereby enriching the understanding of multifaceted efficiency dynamics.

The groundbreaking introduction of the network SBM model in this study facilitates the concurrent assessment of urban governance and ICT efficiency, offering a nuanced comprehension of their intricate interplay and collective impact on urban sustainability. This addresses a conspicuous gap in existing literature where studies often employed traditional DEA models, neglecting the complex relationships between governance, ICT, and sustainability. The incorporation of the network SBM model presents a methodological advancement that significantly refines the precision of efficiency evaluations. The study addresses the research gap by introducing a scoring system using the network SBM of the DEA model. The aim is to identify inefficient DMUs and establish a new urban operational indicator framework. This framework integrates basic indicators of the triple bottom line (economic, social, and environmental sustainability) with governance and ICT indicators to comprehensively address inefficiency (Chang et al. 2013).

This study makes several notable contributions to the field: (1) Filling a gap in China’s urban efficiency indicator system, this study incorporates urban governance, ICT, and sustainability in the DEA-based evaluation framework. (2) The network SBM model enables simultaneous assessment of urban governance and ICT efficiency, providing a precise evaluation of input and output efficiency. (3) Examining 38 major cities in China, this study scrutinized the shifts in relative rankings concerning urban governance, ICT, and sustainability efficiency over five years after the urbanization plan, thereby offering valuable insights into the efficacy of urban reforms. (4) Empirical results from this study provide evidence of differences in urban comprehensive operation, governance, ICT, and sustainability efficiency in coastal and inland areas of China, and shed light on the causes of these disparities.

Literature review

Urban governance

The inception of Local Agenda 21 was a turning point in the development of urban systems and policies, with the aim of promoting sustainable changes through innovative approaches (Raven et al. 2019). The effective planning of urban development by governments is crucial in achieving sustainable development, as policies in areas such as economic development, social welfare, and environmental protection all impact urban sustainability. However, factors such as the centralization of power, political risk, and corruption can impede a city’s ability to implement sustainable development policies (Neirotti et al. 2014). In China, the top-down governance system, with its centralized administrative power, results in cities relying heavily on the leadership of the national government and the support of local governments (Raven et al. 2019). Adequate access to public funds can facilitate the implementation of plans. China’s transition from a regionally decentralized authoritarian (RDA) regime, characterized by political centralization and economic regional decentralization, to a more decentralized system, has allowed local governments to impact regional economic development, environmental protection, and social stability through reforms and policies, under the supervision of the central government (Xu 2011). Urban development strategies in Chinese cities may vary, but all must align with the central government’s requirements.

As the second-largest economy globally, the Chinese government is pursuing the development of smart cities, which can be achieved through the integration of information and communication technology (ICT) (Chourabi et al. 2012). In 2012, the government first proposed the construction of smart cities, and by 2019, it had invested in over 700 projects in more than 500 cities, with 290 smart cities built and evaluated annually (Zhang et al. 2021; Zhu et al. 2019). The central government has invested significant funds in many cities, requested local governments to establish ICT infrastructure, and announced policies focused on environmental protection and ICT (Dameri et al. 2019) to bridge the gaps between different cities.

The literature has established the importance of governance in promoting sustainability. Research by Feng et al. (Feng et al. 2022) indicated a positive relationship between government spending and green economic performance, while Bulkeley and Betsill (Bulkeley and Betsill 2005) illustrated how land use and transportation planning in the UK helped mitigate climate change and reduce energy consumption and waste. Owens and Cowell (Owens and Cowell 2011) suggested the need for changes in traffic road planning and car control to reduce greenhouse gas emissions in the transportation sector.

The chosen variables associated with government efficiency, derived from extant research, encompass a spectrum of dimensions. These comprise general government public expenditures (Fan et al. 2021; Xiao et al. 2021)alongside specific categories of public spending. The delineated areas encompass public finance expenditures related to science and technology, financial allocations for education, disbursements for health-related initiatives, investments in infrastructure, and the quantification of municipal expenditures on a per capita basis (Kairui 2018; Ouertani et al. 2018). In this study, to assess the effectiveness of governance, fiscal revenues, expenditures, employment, and investment in science and technology were adopted as representative indicators, drawing on prior research and empirical evidence.

Urban ICT

Keidanren (2016) posits that the stages of social development span from the hunting society to the agricultural society, industrial society, information society, and finally, the super-intelligent society. The shift from Industry 3.0 to Industry 4.0 marks the transition from a machine-led manufacturing approach to a digital-led approach, driven by the emergence of technologies such as the Internet of Things (IoT) and Machine-to-Machine (M2M) communication (Oztemel and Gursev 2020). To achieve sustainable urban development, a city must pursue smart urban development and informatization (Yigitcanlar et al. 2019). Information and Communication Technology (ICT) has a vital role in transforming a city’s economy, society, and spaces (Hollands 2008) and optimizing urban governance for sustainable development (Wang and Zhou 2022; Yao et al. 2020).

To improve a city’s informatization level, the construction of ICT infrastructure is crucial, as well as the development and utilization of information resources. Informatization has the potential to accelerate the growth of information industries, create employment opportunities, and attract high-tech talent (Harrison et al. 2010). Furthermore, ICT can enhance city management systems, improve the quality of life for urban residents, and optimize service infrastructures such as air pollution monitoring and sewage treatment for environmental governance.

In the realm of current research, variables utilized for gauging ICT or smart cities encompass a diverse set of metrics. These include the count of fixed telephone users, mobile users, indicators for internet accessibility, the prevalence of broadband users, the employed population, the number of information practitioners, the level of broadband network construction, the coverage rate of optical fiber access, per capita postal service volume, business volume of post and telecommunication services and the real GDP per capita within the information industry (ITU 2003; Kairui 2018; Wang and Feng 2015). This study assesses urban ICT efficiency based on the utilization and application of the information industry, the extent of ICT infrastructure construction, and revenue from Telecommunication Services.

Urban sustainability

The Brundtland Report of 1987 emphasized the need for cities to have plans for sustainable development, which has since been widely recognized globally (Bulkeley and Betsill 2005). For sustainable growth, cities must make efficient use of their resources and prioritize practicality (Walter et al. 1992).

China underwent reforms and opened up in 1978, resulting in rapid economic growth but also environmental degradation and increased social pressure (Chen et al. 2016). These issues, commonly referred to as “big city disease,” pose a threat to the development of Chinese cities. In response, the Chinese government made sustainable development a basic international strategy in 1995, seeking to balance economic growth with social and environmental considerations (Lu et al. 2019).

To address the negative impacts of economic growth on cities, China introduced the “National New-Type Urbanization Plan (2014–2020)” in 2014. The plan aimed to promote sustainable cities by setting objectives for resources, environment, and infrastructure and reducing regional development disparities through spatial distribution and scaling (Chen et al. 2016; Wang et al. 2015).

The concept of sustainable development has three main objectives: economic development, environmental sustainability, and social inclusion (Sachs 2012). This aligns with the idea of the triple bottom line (TBL), first introduced by Elkington in 1994, which advocates for companies to balance profitability, social responsibility, and environmental responsibility for sustainability (Elkington 2018). Over time, the TBL has been interpreted in various ways, such as social, environmental, and economic performance, sustainable development, sustainable environment, sustainable communities, and impact on society, the environment, and economic sustainability (Vanclay 2004).

The TBL is a crucial aspect of urban sustainability and is widely used as a guiding principle in evaluating sustainability performance by integrating social, economic, and environmental variables (Chen and Zhang 2020). Governments worldwide use the TBL as part of their comprehensive assessments of their economy, environment, and society, which form the key conceptual requirements of the City Sustainability Index (CSI) (Mori and Christodoulou 2012). The TBL can be used to evaluate not only the overall sustainability of a city but also the sustainability of its urban communities (Berardi 2013).

This study assesses the sustainable development of 38 cities in China based on the TBL of sustainability. In earlier studies, indicators pertinent to urban sustainability were encompassed within a comprehensive set of metrics, spanning both economic and environmental dimensions. These indicators included CO2 emissions, concentrations of PM10, GDP as the anticipated output, the green rate of construction areas, per capita park green area, total energy consumption, water consumption, annual power consumption, unit employees, GDP, and the greening coverage of existing urban areas (Kairui 2018; Li et al. 2016, 2018; Shi et al. 2009). Additionally, various social indicators were considered, with social metrics encompassing the basic old-age insurance coverage ratio, unemployment insurance coverage ratio, and the insurance ratio of medical insurance for urban employees, while economic metrics included the urban unemployment rate (Li et al. 2018). The factors affecting sustainable development are defined as urban economy, social development, and environmental protection. The study evaluates the effectiveness of sustainable development using representative indicators such as per-capita GRP, employment rate, insurance coverage, and inhalable particle concentration.

Data envelopment analysis (DEA)

DEA was first proposed by Charnes et al. (1978), which was also known as a CCR model, in which, the boundary of an efficiency frontier is constructed by linear programming. When compared with this frontier, the relative efficiency value between Decision Making Units (DMUs) can be identified. After 40 years of academic development, DEA has been applied to many studies. To solve the limitations of the CCR model, many different models have been derived from production possibility sets based on different assumptions, such as BCC, NIRS, FDH, etc. (Berardi 2013; Färe et al. 1985; Tulkens 1993). The traditional DEA treats DMU as a black box, which means that only a set of outputs appears after a set of inputs enters the DMU (Tone and Tsutsui 2014). The disadvantage is ignoring the internal structure or linking of the DMUs, causing the input and output to change proportionally (Tone and Tsutsui 2009).

Tone (2001) proposed a model based on the slacks-based measure, which is a non-radial method. Compared with CCR and BCC models, it can directly deal with excessive input and insufficient output. To make DEA modeling more flexible, Färe and Grosskopf (2000) proposed the use of a network model of “links” in the multi-stage input and output progress. In other words, network DEA uses the link variable to consider the internal structure of the DMU, which can further analyze the resource allocation of different stages and processes. Tone and Tsutsui (2009) proposed the network SBM Model based on SBM and network DEA, which can deal with any network structure in which divisions are connected and links are directed. Network SBM can evaluate comprehensive efficiency and increase the function of evaluating the efficiency of each department. They presented efficiency measuring cases of enterprises, hospitals, financial holding companies, etc. to prove the applicability and practicality of the model.

Recent studies have utilized various DEA models to evaluate urban efficiency. Santana et al. (Santana et al. 2014) used the output-oriented DEA-BCC model to compare the sustainable development of BRICS countries. Sueyoshi and Yuan (2015) evaluated the sustainability of economic development and air quality in China using DEA, while Yang et al. (2016) employed an extended urban metabolism framework and DEA model with negative output to evaluate the sustainability of 22 cities in Taiwan. He et al. (2016) used the DEA and ecological efficiency of Jiangsu Province to assess sustainability, while Feng et al. (2017) used DEA to assess the sustainable development efficiency of cities in coal-fired power plant regions. Yu and Zhang (2019) developed a non-convex meta frontier DEA model to measure the energy efficiency of cities in China, and Sun et al. (2020) used a common weight DEA to evaluate the sustainability performance of Southeast Asia. Additionally, Zhang et al. (2020) applied the SBM-DDF model to analyze the green innovation efficiency of Xi’an. Cui and Wang (2023) utilize a model based on the Super-SBM to assess the influence of China’s green finance efficiency on the attainment of its multidimensional green finance objectives, while Miao and Zhou (2023) utilized the Stochastic Frontier Model approach to analyze the economic efficiency of Hohhot-Baotou-Ordos-Yulin urban agglomeration.

Methodology

Research framework

This study proposes a model based on the network SBM model of Kao (2014), which is used to evaluate the efficiency affecting urban governmental departments, information technology, and urban sustainability. The theoretical framework is divided into three stages: Governance, ICT, and Sustainability with 18 indicators. The overall framework is shown in Fig. 1.

Fig. 1: The framework for measuring the comprehensive efficiency of major cities.
figure 1

This diagram illustrates the decomposition of major cities’ comprehensive efficiency into three main components: Governance, ICT (Information and Communication Technology), and Sustainability.

Full size image

The present study aims to evaluate the efficiency of urban sustainability by using a three-stage framework. In the first stage, fiscal revenue and employment are considered as inputs, while fiscal expenditures and capital expenditures for the information and communication technology (ICT) sector are considered as outputs. In the second stage, the input measures are the number of employees and the city’s ICT infrastructure, which includes the number of users of telephones and Internet services. The output revenue generated in this stage is used as the economic input in the third stage, which focuses on the sustainability inputs of water supply, annual electricity consumption, and total gas supply. These inputs represent the resource and energy use of the city. In the sustainability stage, economic, social, and environmental performance indicators are used to measure the efficiency of the city’s sustainability, as represented by the three aspects of the triple bottom line of sustainability. The economic indicators are gross regional product (GRP) and unemployment; the social indicators are three kinds of social insurance, while the environmental indicators are PM10 and green covered areas. It is important to note that the employment rate and PM10 are negative outputs.

To comprehensively analyze the indicators used to assess the basic characteristics of urban construction, Tables 1 and 2 were compiled, which include data collected for each indicator over a five-year period, amounting to 190 cases. The output indicator values, excluding negative indicators, have shown an upward trend in the selected urban areas. Typically, cities with higher indicator values, excluding negative indicators, are first-tier cities with strong overall economic development, such as Shanghai, Beijing, and Shenzhen. These cities have prioritized sustainable development, resulting in stable negative indicator values despite the urban expansion. Electricity consumption, which is influenced by factors such as overall economic development, key electricity-consuming industries, geography, and climate, exhibits the greatest variation among the indicators and can serve as an indirect measure of urban development. For example, Shanghai, as a leading commercial center, had the highest electricity consumption of 1568,5775GWh in 2019, while Suzhou ranked second with 1544,4800GWh due to its thriving industrial sector (NBS 2020).

Table 1 Symbols and definitions of the indicators.
Full size table
Table 2 Descriptive statistics of the indicators.
Full size table

Network performance evaluation model

This study evaluated 38 cities, which were used as the evaluation decision-making units (DMUs), then DMU (j = 1, 2, …, n). At the stage of governance efficiency, each \({{\rm{DMU}}}_{j}\) has two input items \({x}_{r0}^{g}(r=1,2)\) and produces link activity \({Z}_{j}^{{gi}}\) to the ICT efficiency stage and link activity \({Z}_{j}^{{gs}}\) in the sustainability efficiency stage. At the stage of ICT efficiency, each \({{\rm{DMU}}}_{j}\) has three input variables \({x}_{t0}^{i}(t=1,2\) produce a link activity \({Z}_{j}^{{is}}\). At the stage of sustainability efficiency, each \({{\rm{DMU}}}_{j}\) has three input items \({x}_{p0}^{s}(p=1,2,3)\) produce five output items \({y}_{{qj}}^{s}\)(q = 1, 2…5) and two undesirable output items \({u}_{k0}^{s}(k=1,2)\). The comprehensive efficiency score (\({\sigma }_{0}^{* }\)) of \({{\rm{DMU}}}_{0}\) based on the network DEA-SBM model can be defined as:

$${\rm{Min}}{\sigma }_{0}^{* }=\frac{{w}_{g}\left[1-\frac{1}{2}\left({\sum }_{r=1}^{2}\frac{{{xs}}^{g}}{{x}_{r0}^{g}}\right)\right]+{w}_{i}\left[1-\frac{1}{3}\left({\sum }_{t=1}^{3}\frac{{{xs}}^{i}}{{x}_{t0}^{i}}\right)\right]+{w}_{s}\left[1-\frac{1}{5}\left({\sum }_{p=1}^{3}\frac{{{xs}}^{s}}{{x}_{p0}^{s}}+{\sum }_{k=1}^{2}\frac{{b}_{k}^{s}}{{u}_{k0}^{s}}\right)\right]}{{w}_{g}\left[1+\frac{1}{2}\left(\frac{{{sz}}^{g}}{{Z}_{0}^{{gi}}}+\frac{{{ez}}^{g}}{{Z}_{0}^{{gs}}}\right)\right]+{w}_{i}\left[1+\left(\frac{{{ez}}^{i}}{{Z}_{0}^{{is}}}\right)\right]+{w}_{s}\left[1+\frac{1}{5}\left({\sum }_{q=1}^{5}\frac{{s}_{q}^{s}}{{y}_{q0}^{s}}\right)\right]}$$
(1)

subject to

$${x}_{r0}^{g}=\mathop{\sum }\limits_{j=1}^{n}{x}_{{rj}}^{g}{\lambda }_{j}^{g}+{{xs}}^{g}$$
(1.1)
$${Z}_{0}^{{gi}}=\mathop{\sum }\limits_{j=1}^{n}{Z}_{j}^{{gi}}{\lambda }_{j}^{g}-{{sz}}^{g}$$
(1.2)
$$\mathop{\sum }\limits_{j=1}^{n}{Z}_{j}^{{gi}}{\lambda }_{j}^{g}=\mathop{\sum }\limits_{j=1}^{n}{Z}_{j}^{{gi}}{\lambda }_{j}^{i}$$
(1.3)
$${Z}_{0}^{{gi}}=\mathop{\sum }\limits_{j=1}^{n}{Z}_{j}^{{gi}}{\lambda }_{j}^{o}-{{sz}}^{i}$$
(1.4)
$${Z}_{0}^{{gs}}=\mathop{\sum }\limits_{j=1}^{n}{Z}_{j}^{{gs}}{\lambda }_{j}^{g}-{{ez}}^{g}$$
(1.5)
$$\mathop{\sum }\limits_{j=1}^{n}{Z}_{j}^{{gs}}{\lambda }_{j}^{g}=\mathop{\sum }\limits_{j=1}^{n}{Z}_{j}^{{gs}}{\lambda }_{j}^{s}$$
(1.6)
$${Z}_{0}^{{gs}}=\mathop{\sum }\limits_{j=1}^{n}{Z}_{j}^{{gs}}{\lambda }_{j}^{s}-{{ez}}^{s}$$
(1.7)
$${x}_{t0}^{i}=\mathop{\sum }\limits_{j=1}^{n}{x}_{j}^{i}{\lambda }_{j}^{i}+{{xs}}^{i}$$
(1.8)
$${Z}_{0}^{{is}}=\mathop{\sum }\limits_{j=1}^{n}{Z}_{j}^{{is}}{\lambda }_{j}^{i}-{{sz}}^{i}$$
(1.9)
$$\mathop{\sum }\limits_{j=1}^{n}{Z}_{j}^{{is}}{\lambda }_{j}^{i}=\mathop{\sum }\limits_{j=1}^{n}{Z}_{j}^{{is}}{\lambda }_{j}^{s}$$
(1.10)
$${Z}_{0}^{{is}}=\mathop{\sum }\limits_{j=1}^{n}{Z}_{j}^{{is}}{\lambda }_{j}^{s}-{{sz}}^{s}$$
(1.11)
$${x}_{p0}^{s}=\mathop{\sum }\limits_{j=1}^{n}{x}_{j}^{s}{\lambda }_{j}^{s}+x{s}^{s}$$
(1.12)
$${y}_{q0}^{s}=\mathop{\sum }\limits_{j=1}^{n}{y}_{{qj}}^{s}{\lambda }_{j}^{s}-{s}_{q}^{s}$$
(1.13)
$${u}_{0}^{s}=\mathop{\sum }\limits_{j=1}^{n}{u}_{j}^{s}{\lambda }_{j}^{s}+{b}^{s}$$
(1.14)
$$\mathop{\sum }\limits_{j=1}^{n}{\lambda }_{j}^{g}=1$$
(1.15)
$$\mathop{\sum }\limits_{j=1}^{n}{\lambda }_{j}^{i}=1$$
(1.16)
$$\mathop{\sum }\limits_{j=1}^{n}{\lambda }_{j}^{s}=1$$
(1.17)
$${{xs}}^{g},{{sz}}^{g},{{sz}}^{i},{{ez}}^{g},{{ez}}^{s},{{xs}}^{i},{{sz}}^{i},{{sz}}^{s},x{s}^{s},{s}_{q}^{s}{,b}^{s},{\lambda }^{g},{\lambda }^{i},{\lambda }^{s}\ge 0$$

The \({\lambda }_{j}^{g}\) (j = 1, 2, …, n), \({\lambda }_{j}^{i}\) (j = 1, 2, …, n) and \({\lambda }_{j}^{s}\) (j = 1, 2, …, n) are the intensity variables of the jth city in constructing the Governance, ICT, and sustainability efficiency frontiers. The weights of the three efficiency stages respectively denoted by \({w}_{g},{w}_{i}{\rm{and}}\) \({w}_{s}\), represent the relative significance and contribution. In this study, they satisfy \({w}_{g}+{w}_{o}+{w}_{i}=1.\) The \({{xs}}^{g},{{xs}}^{i},{and\; x}{s}^{s}\) represent the excess input in the efficiency stages of governance, ICT, and sustainability. The \({{sz}}^{g}{{ez}}^{g}{{ez}}^{i}\) respectively denotes the shortage of the link activity in the governance and ICT efficiency stage, \({s}_{q}^{s}\) denotes the shortage of the qth output in sustainability efficiency, and \({b}^{s}\) denotes the excess of the undesirable output in the sustainability efficiency stage.

The calculation of governance efficiency is subject to Eqs. (1.1)–(1.7); The calculation of ICT efficiency is subject to Eqs. (1.8)–(1.11); The calculation of Sustainable efficiency is subject to Eqs. (1.12)–(1.14). Adding Eqs. (1.15)–(1.17) allow the returns to scale to be variable.

Decompose \({{\rm{DMU}}}_{0}\) into governance efficiency score (\({\sigma }_{0}^{{\rm{g}}* }\)), ICT efficiency score (\({\sigma }_{0}^{{\rm{i}}* }\)) and sustainability efficiency score (\({\sigma }_{0}^{{\rm{s}}* }\)) using the optimal results (\({{xs}}^{g* },{{sz}}^{g* },{{sz}}^{i* },{{ez}}^{g* },{{ez}}^{s* },{{xs}}^{i* },{{sz}}^{i* },{{sz}}^{s* },x{s}^{s* },{s}_{q}^{s* }{,b}^{s* },{\lambda }^{g* },{\lambda }^{i* },{\lambda }^{s* }\)) in Eq. (1) as follows:

$${\rm{Minimize}}\,{\sigma }_{0}^{{\rm{g}}* }=\frac{1-\frac{1}{2}\left({\sum }_{r=1}^{2}\frac{{{xs}}^{g}}{{x}_{r0}^{g}}\right)}{1+\frac{1}{2}\left(\frac{{{sz}}^{g}}{{Z}_{0}^{{gi}}}+\frac{{{ez}}^{g}}{{Z}_{0}^{{gs}}}\right)}$$
(2)
$${\rm{Minimize}}\,{\sigma }_{0}^{{\rm{i}}* }=\frac{1-\frac{1}{3}\left({\sum }_{t=1}^{3}\frac{{{xs}}^{i}}{{x}_{t0}^{i}}\right)}{1+\left(\frac{{{ez}}^{i}}{{Z}_{0}^{{is}}}\right)}$$
(3)
$${\rm{Minimize}}\,{\sigma }_{0}^{{\rm{s}}* }=\frac{1-\frac{1}{5}\left({\sum }_{p=1}^{3}\frac{{{xs}}^{s}}{{x}_{p0}^{s}}+{\sum }_{k=1}^{2}\frac{{b}_{k}^{s}}{{u}_{k0}^{s}}\right)}{1+\frac{1}{5}\left({\sum }_{q=1}^{5}\frac{{s}_{q}^{s}}{{y}_{q0}^{s}}\right)}$$
(4)

Sources of samples

Golany and Roll (1989) suggested that when using the Data Envelopment Analysis (DEA) model, the number of observations should not be less than twice the sum of inputs and outputs. Based on this recommendation, the sample selection in this study consists of the top 38 cities in the 2020 urban business charm ranking, released by New First-tier Cities Research, which includes 4 first-tier cities, 15 new first-tier cities, and 19 second-tier cities. The observation period for this study ranges from 2015 to 2019. The data used in this study are derived from various official sources, including the China City Statistical Yearbook (2016–2020), China Statistical Yearbook (2016–2020), the Yearbook of China Communication (2016–2020), and statistical yearbooks of each province and city.

In 2014, the State Council of the People’s Republic of China defined major cities as those with a population of over 1 million (The State Council 2020). The largest city in terms of population among the selected cities was Chongqing with 34.16 million people, while the smallest was Dongguan with 2.51 million people in 2019. The total population of the 38 selected cities is approximately 313.28 million, accounting for approximately 22% of the national population. These cities possess favorable economic conditions and infrastructure, which attract large numbers of migrant workers. As highly populated areas, they are confronted with various challenges such as social contradictions, environmental pollution, and traffic congestion. However, the selected cities are widely dispersed across China, which results in significant variations in geographical environments and natural resources, economic structures, and population densities. Therefore, the efficiency of urban sustainability is influenced by a range of factors that vary across the selected cities.

Empirical results

Efficiency value analysis

The study employed the network SBM model to analyze the overall efficiency of 38 cities in China from 2015 to 2019. Figure 2 presents a visual representation of this analysis, providing a comprehensive perspective on performance across various spatial DMUs or over time. The illustration comprises five maps, each corresponding to a specific year. Serial numbers are assigned to the positions of 38 cities on each map, and blue indicators signify the overall efficiency of each city. The diverse shades of color convey distinct levels of overall efficiency, with dark blue indicating high efficiency and light blue indicating low efficiency. Positioned on the right side of each map is a color legend elucidating the relationship between different color shades and their corresponding levels of overall efficiency. Across all five maps, a discernible gradient is observed from light blue to dark blue, particularly from inland to coastal areas, indicating a notable difference in efficiency between coastal and inland cities.

Fig. 2: Spatial distribution of comprehensive efficiency in 38 cities during 2015–2019.
figure 2

This figure demonstrates the spatial distribution and efficiency scores of 38 cities, using a color gradient to signify varying levels of comprehensive efficiency, and reveals that coastal cities consistently outperform inland cities from 2015 to 2019.

Full size image

The meticulous presentation of detailed efficiency values for each Decision Making Unit (DMU) in Table 3 offers a nuanced perspective on the comprehensive operational efficiency and specific stage efficiencies (governance, ICT, and sustainability) across 38 Chinese cities from 2015 to 2019. In this table, a score of 1 signifies optimal comprehensive efficiency or stage efficiency in the cities. Notably, the most substantial enhancement in urban efficiency occurred in the sustainability stage, with the number of effective cities rising from 25 in 2015 to 29 in 2019. On average, coastal cities surpass their inland counterparts in overall efficiency, reflected in their higher scores.

Table 3 Overall and divisional efficiencies for cities in China.
Full size table

As indicated by Table 3, specific cities have consistently held a relative ranking within the lower 15% over the last five years, like Wuhan, Tianjin, and Shijiazhuang. Conversely, certain cities have demonstrated considerable enhancements in their relative rankings during this five-year period, exemplified by Harbin’s ascent from the 33rd position in 2015 to the 16th in 2019. Nevertheless, contrasting trends are observable as well, with cities like Shenzhen experiencing a decline, transitioning from the top position in 2015 to the 18th in 2019.

Changzhou and Jiaxing consistently emerge as benchmark cities in the annual comparative analysis of comprehensive efficiency from 2015 to 2019, showcasing an exceptional ability to optimally allocate and utilize resources. The governments of Changzhou and Jiaxing strategically leverage their geographical advantages to foster the development of characteristic industries. Furthermore, they prioritize the judicious use of financial funds to support coordinated urban resource management and environmental development. For instance, Changzhou strategically shifted its economic structure from a reliance on the secondary industry to a focus on the tertiary industry in 2015, concurrently committing to the ambitious goal of building an eco-friendly city. Jiaxing, on the other hand, has been consistently enhancing its Information and Communication Technology (ICT) infrastructure, integrating Internet elements into the urban landscape and gaining international recognition by hosting the World Internet Conference since 2014.

Beijing and Shanghai, as the largest cities in China, exhibit efficiency values close to 1, indicating highly effective overall operating efficiency, with minor adjustments in resource input required. Examining governance efficiency, Beijing and Shanghai consistently exhibit excellence, while some cities like Guangzhou and Dongguan display fluctuations in this dimension, contributing to overall disparities. In the field of ICT efficiency, Shenzhen and Beijing are consistent leaders, although certain inland cities like Changchun and Jiaxing demonstrate competitive ICT efficiencies. Generally, coastal cities tend to exhibit superior ICT efficiencies, influencing their overall standings. Sustainability efficiency reveals variations among cities, with Dongguan and Changzhou consistently attaining top scores, while inland cities such as Shijiazhuang face sustainability challenges, evident in their lower scores.

Slack analysis for improvements

To further analyze the efficiency of cities and provide improvement suggestions, this study gives suggestions for improvement to inefficient cities, this study analyzed the slack in input and output usage of each city in 2019 to obtain the potential improvement averages. Table 4 shows the results obtained by dividing the slack by the original values of the input, output, and intermediate products of each city. The values in the table represent the input proportion that should be reduced and the output proportion that should be increased. Take Shijiazhuang with the lowest score for overall efficiency as an example, the part of its input \({x}_{1j}^{g},{x}_{2j}^{g},{x}_{1j}^{i},{x}_{2j}^{i},{x}_{3j}^{i},{x}_{1j}^{s},{x}_{2j}^{s},{x}_{3j}^{s}\) should be reduced, respectively, 31.96%, 32.88%, 68.1%, 50.38%, 50.46%, 11.46%, 24.46%, 66.84%.The part of its output \({y}_{1j}^{s},{y}_{2j}^{s},{y}_{3j}^{s},{y}_{4j}^{s},{y}_{5j}^{s}\) should be increased, respectively, 100.98%, 0, 18.18%, 55.92%, 0. The part of its undesirable output \({u}_{1j}^{s}{,u}_{2j}^{s}\) should be reduced, respectively, by 41.8% and 53.39%.

Table 4 Potential improvement for the 38 cities.
Full size table

Regarding urban governance, some essential inputs cannot be omitted, such as government revenue (\({x}_{1j}^{g}\)) and ICT infrastructure (\({x}_{2j}^{i},{x}_{3j}^{i}\)). A larger input value indicates that the city has invested more resources in the construction of infrastructure. However, some inputs, such as the number of public management staff (\({x}_{2j}^{g}\)), can be appropriately reduced. In the sustainability stage, to reduce the water and electricity consumption of urban residents (\({x}_{1j}^{s},{x}_{2j}^{s}\)), the government should advocate resource conservation for urban residents, promote the use of clean energy, and gradually reduce the use of gas. Therefore, in order to improve the overall operation governance efficiency of the city, it is necessary to increase GRP(\({y}_{1j}^{s}\)), improve the coverage of social welfare insurance (\({y}_{2j}^{s},{y}_{3j}^{s},{y}_{4j}^{s}\)) and urban greening coverage output (\({y}_{5j}^{s}\)), while reducing the negative output of the employment rate and inhalable particles\(({u}_{1j}^{s}{,u}_{2j}^{s}\)).

The effect of region on efficiencies

In the context of China’s reform and opening up, the growth of urban conglomerates in the Yangtze River Delta and Pearl River Delta regions has driven the development of nearby cities, resulting in a geographical disparity between coastal and inland regions. To investigate the potential difference in comprehensive efficiency between these two regions, this study proposes the following research hypotheses: From 2014 to 2019, there was no difference between coastal and inland cities in comprehensive operation efficiency (\({H}_{a}\)), governance efficiency (\({H}_{b}\)), ICT efficiency (\({H}_{c}\)) or sustainability efficiency (\({H}_{d}\)). The Kruskal–Wallis test was employed to evaluate the significance level of 5% to compare the efficiency scores of comprehensive operation, governance, ICT, and sustainability based on the geographical location of the cities (coastal or inland).

According to the geographical location, the 38 cities are divided into coastal and inland areas, including 17 coastal cities and 21 inland cities. The results of the Kruskal–Wallis test are shown in Table 5. It is evident that \({H}_{a}{H}_{d}\) showed significant differences in 2015; \({H}_{a}{{H}_{b}{H}_{c}H}_{d}\) exhibited significant differences in 2016; \({H}_{a}{{H}_{b}H}_{c}\) revealed significant differences in 2017. However, in 2018, \({H}_{a}{{H}_{b}{H}_{c}H}_{d}\) did not show significant differences. Intriguingly, \({H}_{a}{{H}_{b}H}_{c}\) again revealed significant differences in 2019.The findings of the Kruskal–Wallis test revealed significant differences in the comprehensive efficiency rates of coastal and inland cities from 2015 to 2017, as indicated by the statistical results.

Table 5 The results of the Kruskal–Wallis test.
Full size table

Figure 3 presents the 5-year average overall efficiency of coastal and inland cities. The data indicates that from 2015 to 2019, although the gap between inland and coastal cities displayed a decreasing trend in the initial four years, it expanded in the fifth year. Average overall efficiency scores for coastal cities consistently exceeded 0.8, while those for inland cities struggled to surpass the 0.8 threshold. Importantly, the efficiency gap between inland and coastal cities during the sustainability phase was significant in 2015–2016 but became non-significant from 2017 onwards.

Fig. 3: Comprehensive efficiency of coastal and inland cities during 2015–2019.
figure 3

This bar graph compares the average comprehensive efficiency of coastal and inland areas from 2015 to 2019, clearly illustrating higher efficiency scores for coastal cities each year.

Full size image

In the governance dimension, the average efficiency scores for inland cities from 2015 to 2019 were 0.662, 0.597, 0.652, 0.628, and 0.593, compared to coastal cities, which scored 0.780, 0.753, 0.808, 0.781, and 0.771. Regarding the ICT dimension, the average efficiency scores for inland cities over the same period were 0.762, 0.739, 0.791, 0.822, and 0.787, while coastal cities scored 0.871, 0.873, 0.889, 0.891, and 0.872.

It is evident that there is a substantial efficiency gap between inland and coastal cities in both governance and ICT. In conclusion, the overarching efficiency disparity between these two categories of cities can be predominantly attributed to variations in governance and ICT.

The Chinese government has acknowledged the necessity for balanced development between coastal and inland regions, implementing various policies to address this imperative. The “Urbanization Plan” initiated in 2014 aimed to balance the development of national land space, resulting in a gradual and orderly transfer of industries from coastal and international areas to urban agglomerations in the central and western regions of China. Inland cities such as Chengdu and Chongqing, and the middle reaches of the Yangtze River, have leveraged this initiative to undertake the process of new industrialization and enhance their infrastructure. Furthermore, the “Vision and Actions on Jointly Building Silk Road Economic Belt and 21st-Century Maritime Silk Road” plan, launched in 2015, not only prioritizes economic and urban development but also highlights the pivotal role of digital infrastructure in enhancing interconnectivity. This strategic initiative has spurred economic advancement across 18 provinces in China, fostering connections among provinces, cities, and nations spanning Asia, Africa, and Europe. By successfully propelling economic development along the Belt and Road, the plan has broadened the global reach and accessibility of inland cities (Xie et al. 2023). Despite the active efforts of inland cities to narrow the development gap, the economic prowess and well-established infrastructure of coastal cities persist as noteworthy factors that cannot be overlooked. The optimal scenario envisions concurrent progress for both coastal and inland cities.

Conclusions and policy implications

This study presents a novel model for evaluating the overall operational efficiency of urban areas using network SBM methodologies. The model decomposes comprehensive efficiency into three key elements: governance, ICT, and sustainability. The model is applied to assess the efficiency of 38 significant cities in China using data collected from 2015 to 2019. The primary findings of this study are as follows:

  1. 1.

    A majority of the cities lacked efficiency in governance and ICT. The indicators of governance employment and mobile phone usage were particularly inefficient, indicating challenges with financial structure and over-employment in urban governments. Additionally, the government’s investment in ICT infrastructure did not necessarily lead to efficient economic outcomes.

  2. 2.

    In the context of China’s reform and opening up, the growth of urban conglomerates in the Yangtze River Delta and Pearl River Delta regions has led to a geographical disparity between coastal and inland areas. The Kruskal–Wallis test was used to compare efficiency scores, dividing 38 cities into coastal and inland. Results show significant differences in comprehensive efficiency from 2015 to 2017, with the gap narrowing initially but expanding in 2019. Coastal cities consistently outperformed inland cities, especially in governance and ICT.

  3. 3.

    The study discerned that the overall operational efficiency of 14 cities, each with an urban population surpassing 10 million, proved suboptimal in 2019, registering an average efficiency score of merely 0.7168. In stark contrast, 7 out of the 24 cities with a population below 10 million demonstrated superior overall operational efficiency, boasting an average score of 0.8637. The detailed breakdown of cities with a permanent population exceeding 10 million is presented in Table 3. This observation underscores that cities characterized by larger urban populations contend with heightened ecological and social pressures, rendering sustainable development more intricate and resulting in comparatively diminished overall efficiency.

  4. 4.

    To enhance the overall operational efficiency of cities, it is imperative to make strategic adjustments to the economic structure and foster the development of green industries. The application of Network SBM methodologies facilitates inefficient cities in identifying benchmark cities, learning from them, drawing inspiration, and subsequently refining urban planning strategies. Notably, among the 14 cities identified by the model construction with an overall efficiency score of 1, 11 display a predominant focus on developing the tertiary industry. This emphasis is substantiated by data from the yearbook, revealing that the proportion of the tertiary industry in these cities’ regional GDP surpasses that of the primary and secondary industries. The cities exhibiting this focus include Qingdao, Kunming, Dalian, Harbin, Jinan, Wenzhou, Nanning, Changchun, Shijiazhuang, Jinhua, and Changzhou. A case in point is Qingdao, which illustrates a five-year period marked by rapid and effective industrial structure optimization. The proportional distribution of primary, secondary, and tertiary industries concerning Gross Regional Product (GRP) shifted from 3.9:43.3:52.8 in 2015 to a more harmonized 3.49:35.63:60.89 in 2019. Consequently, the overall efficiency, computed within the specified model architecture, progressed from 0.80181 in 2015 to a perfect score of 1 in 2019. The case study of Qingdao illustrates the positive impact of policies promoting the tertiary and green industries, indicating a symbiotic relationship between industrial structure optimization and increased overall efficiency. Therefore, cities facing inefficiencies can propel sustainable urban development through policies directed at fostering the development of tertiary and green industries.

While it is acknowledged that other short-term and conventional policies may impact urban development efficiency, the commencement of the “National New-Type Urbanization Plan” in 2014 stands as a pivotal moment. This initiative sparked substantial reforms in city governance, economic restructuring, and a pronounced focus on environmental protection. A discernible outcome of these reforms is the notable progress in sustainability practices, evident in a substantial increase in the number of cities implementing efficient sustainability measures from 2015 to 2019. This dynamic reflects a pervasive upward trajectory in the sustainability efficiency of Chinese cities.

A geographical disparity between coastal and inland areas becomes evident, with coastal cities consistently outperforming their inland counterparts. The study nuances this gap, emphasizing its initial narrowing and subsequent expansion from 2015 to 2019. Despite earnest endeavors by the Chinese government to foster balanced development, coastal cities continue to wield significant economic strength. The imperative for both coastal and inland cities to progress collectively for optimal development is underscored. Key government initiatives, such as the “Urbanization Plan” and the “Belt and Road” initiative, have contributed to shaping this landscape. However, challenges persist in achieving equilibrium in development.

The study identifies prevalent inefficiencies in governance and ICT, emphasizing challenges in financial structures and potential over-employment in urban government bodies. Consequently, the government is urged to address these challenges by streamlining bureaucracy, restructuring financial expenditures, and enhancing governance efficiency. Furthermore, local governments are urged not to overlook the potential of ICT and informatization. Increased investment in science and technology is essential to meet the smart city requirements mandated by the central government. To propel a smarter, low-carbon urban economy, local governments should focus on developing robust ICT infrastructure and seamlessly integrating ICT technology with traditional industries. Additionally, transforming the industrial structure and promoting green industries require concerted efforts to reduce the consumption of ecological resources, enforce strict controls on factory emissions of sewage and waste gas, and enhance social security, basic life services, air quality monitoring, green spaces, and residents’ living environments.

In conclusion, achieving sustainable urban development demands coordinated efforts across various domains, including ICT, the economy, society, and the environment. While acknowledging the significant strides made by the Chinese government in promoting sustainable urbanization, the study emphasizes the need for ongoing efforts to mitigate the negative impacts of urban development and ensure the enduring sustainability of cities. Thus, governance, ICT, and sustainability emerge as pivotal facets in effectively addressing the multifaceted challenges inherent in urban development.